Columnar structured FePt films epitaxially grown on large lattice mismatched intermediate layer

The microstructure and magnetic properties of the FePt films grown on large mismatched ZrN (15.7%) intermediate layer were investigated. With using ZrN intermediate layer, FePt 10 nm films exhibited (001) texture except for some weaker FePt (110) texture. Good epitaxial relationships of FePt (001) <100>//ZrN (001) <100>//TiN (001) <100> among FePt and ZrN/TiN were revealed from the transmission electron microscopy (TEM) results. As compared with TiN intermediate layer, although FePt-SiO2-C films grown on ZrN/TiN intermediate layer showed isotropic magnetic properties, the large interfacial energy and lattice mismatch between FePt and ZrN would lead to form columnar structural FePt films with smaller grain size and improved isolation. By doping ZrN into the TiN layer, solid solution of ZrTiN was formed and the lattice constant is increased comparing with TiN and decreased comparing with ZrN. Moreover, FePt-SiO2-C films grown on TiN 2 nm-20 vol.% ZrN/TiN 3 nm intermediate layer showed an improved perpendicular magnetic anisotropy. Simultaneously, columnar structure with smaller grain size retained.

, 3.44% for KTaO 3 15 , 1.44% for SrTiO 3 10-15 and − 1.63% for LaAlO 3 12,15 , respectively. It is found the lattice mismatch plays a crucial role in the growth of epitaxial thin films 16,17 . Y. Chen 18 et al. had proposed that the mismatched film preferred island structure when the size of the mono epilayer exceeded the critical size, which is inversely proportional to the mismatch strain ε xx 6 . Therefore, larger lattice mismatch (larger than TiN) will be favor for the FePt films with smaller grain size. However, the large mismatch may not be favored for good (001) texture and perpendicular anisotropy. Until now, the effects of larger lattice mismatch (11-15%) on the magnetic properties and microstructure of FePt films have not been reported and the epitaxial growth correlation between FePt and larger mismatched intermediate layer still remains unclear. In this paper, we deposited FePt thin films on ZrN intermediate layer to systematically investigate the larger lattice mismatch effects on the microstructure and magnetic properties of FePt films, and the lattice mismatch between FePt films and ZrN layer is 15.7%. Figure 1(a) shows XRD 2θ spectra of FePt 10 nm films grown on different intermediate layer. It can be seen that the FePt films grown on TiN intermediate layer exhibited (001) preferred orientation. This is due to the epitaxial growth of the FePt on the (200) textured TiN intermediate layer 19,20 . For FePt films grown on ZrN intermediate layer, in addition to the FePt (001) and (002) peaks, FePt (110) peaks were also observed. As compared with the intensities of FePt (001) and (002) peaks, the intensity of FePt (110) peak was much weaker, indicating the (001) texture was dominant in this FePt film. However, No peaks ZrN were shown in Fig. 1(a), which was due to the small quantity of ZrN. Although both FePt films revealed perpendicular anisotropy, the FePt films grown on ZrN/ Scientific RepoRts | 6:34637 | DOI: 10.1038/srep34637

Results
TiN intermediate layer (Fig. 1b) showed a bigger open-up in in-plane loops than that of using TiN intermediate layer ( Fig. 1c) Fig. 1(d,e). As seen, island growth of FePt grains with distinct grain boundary was formed for both intermediate layers. Moreover, the grain size of FePt grown on ZrN/TiN was smaller than that of grown on TiN intermediate layer.
The cross section TEM results of FePt 10 nm grown on TiN intermediate layer had been reported previously (no shown here) and good epitaxial relationship of FePt (001) < 100> //TiN (001) < 100> had been confirmed 20 . The in-plane mismatch ε = (a sub − a FePt )/a sub between TiN and FePt was around 9.1%. In order to investigate the epitaxial growth relationship between FePt and ZrN, the cross-sectional TEM were carried out.

Discussion
Based on the results above, in order to further study the effect of large lattice mismatch on microstructure and magnetic properties of FePt films, FePt 4 nm-SiO 2 35 vol.%-C 20 vol.% films were fabricated on these intermediate layer.   Figure 4 illustrates the microstructure and magnetic properties of FePt 4 nm-SiO 2 35 vol.%-C 20 vol.% films with different intermediate layers. As seen from the planar view TEM images (Fig. 4a), the grain shape of FePt films grown on TiN intermediate layer was maze-like. With using ZrN/TiN intermediate layer, the grain boundaries became more distinct, and the grain changed from maze-like to equiaxed shape (Fig. 4b). Furthermore, the grain size was reduced from 11.2 ± 3.6 nm to 8.4 ± 1.8 nm with an improved grain size distribution, which was consistent with the assumption above-larger lattice mismatch (larger than TiN) will be favor for the FePt films   (Fig. 4c) and well isolated one layer columnar structure (Fig. 4f). Moreover, the perpendicular magnetic anisotropy was improved. The perpendicular coercivity H c⊥ increased from 10.6 kOe with TiN intermediate layer to 14.7 kOe with TiN 2 nm-20 vol.% ZrN/TiN 3 nm intermediate layer (Fig. 4i). The saturated magnetization M s of all samples were around 750 emu/cc and the M s of FePt film was a little larger than that of FePt films grown on ZrN intermediate layer.
In order to further investigate the microstructure evolution of FePt with different intermediate layers, high resolution TEM was carried out. The results are shown in Fig. 5. FePt films grown on a pure TiN intermediate layer exhibited semispherical grains with a contact angle smaller than 90° (Fig. 5a), whereas the film grown on ZrN/TiN showed rectangular grains with a contact angle of around 90° (Fig. 5b). According to Young's equation, (where γ s, γ f and γ fs are the surface energy of the substrate, the surface energy of the films, and the interfacial energy between the substrate and the films, respectively), θ is the contact angle, a large contact angle corresponded to the small surface energy of the substrate and the large interfacial energy. The surface energy of TiN and ZrN are γ TiN ≈ 1.28 J/m 2 and γ ZrN ≈ 1.44 J/m 2 22 , respectively. As γ ZrN is larger than γ TiN , thus the large contact angle obtained by using ZrN intermediate layer was caused by the large interfacial energy between FePt and ZrN, which would promote island growth and thus good grain isolation. Although the larger surface energy of ZrN was not benefit for island growth of FePt films, the smaller grain size with good grain isolation and     Fig. 5(c), (001) and (111) axis of L1 0 FePt aligned very well along ZrTiN (002) and (111) axis, respectively, confirming the epitaxial relationship of FePt (001) < 100> //ZrTiN (001) < 200> . These results confirmed that doping 20 vol.% ZrN into the TiN layer would increase the interfacial energy and the lattice constant, which would be benefit to get the large contact angle and promote island growth and thus obtain good grain isolation. Simultaneously, the optimal lattice constant with 20 vol.% ZrN doping would maintain good epitaxial growth between FePt and ZrTiN and do not deteriorate the perpendicular anisotropy. More than that, the improved isolation of FePt grains would enhance the perpendicular anisotropy of FePt films and make the magnetic properties better than that of using TiN intermediate layer (Fig. 4i). Further increase in the ZrN doping concentration to 40 vol.% caused the deterioration of the isolation and the perpendicular anisotropy (perpendicular coercivity H c⊥ ≈ 13.5 kOe) of FePt-SiO 2 -C films as shown in Fig. 6.

Methods
FePt films fabrication. FePt  Characterization of FePt films. The crystallographic texture was examined with X-ray diffraction (XRD) using Cu K α radiation. The microstructures of the films were characterized by JEOL 2010F transmission electron microscopy (TEM). The morphologies of the samples were examined by Zeiss Supra 40 FE scanning electron microscopy (SEM). The magnetic properties were measured using the alternating gradient force magnetometer (AGFM) at a maximum applied field of 20 kOe at room temperature and using the superconducting quantum interference device (SQUID) at a maximum applied field of 60 kOe at room temperature.